YOLOv10融合DASI，一种新的模块

最新推荐文章于 2025-01-19 00:22:56 发布

weixin_45734021

最新推荐文章于 2025-01-19 00:22:56 发布

阅读量485

点赞数 12

文章标签： YOLO

本文链接：https://blog.youkuaiyun.com/weixin_45734021/article/details/142206239

版权

YOLOv10融合DASI，一种新的模块
yaml文件如下：

# Ultralytics YOLO 🚀, AGPL-3.0 license
# YOLOv10 object detection model. For Usage examples see https://docs.ultralytics.com/tasks/detect

# Parameters
nc: 80 # number of classes
scales: # model compound scaling constants, i.e. 'model=yolov10n.yaml' will call yolov10.yaml with scale 'n'
  # [depth, width, max_channels]
  n: [0.33, 0.25, 1024]

backbone:
  # [from, repeats, module, args]
  - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2      # 64 320 320
  - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4     # 128 160 160
  - [-1, 3, C2f, [128, True]]   #2             # 128 160 160
  - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8     # 256 80 80
  - [-1, 6, C2f, [256, True]]    #4           # 256 80 80
  - [-1, 1, SCDown, [512, 3, 2]] # 5-P4/16  # 512 40 40
  - [-1, 6, C2f, [512, True]]       #6        # 512 40 40
  - [-1, 1, SCDown, [1024, 3, 2]] # 7-P5/32 # 1024 20 20
  - [-1, 3, C2f, [1024, True]]       #8       # 1024 20 20
  - [-1, 1, SPPF, [1024, 5]] # 9            # 1024 20 20
  - [-1, 1, PSA_test, [1024]] # 10               # 1024 20 20

# YOLOv10.0n head
head:
  # 1024 20 20    #512 40 40    #256 80 80
  - [[10, 6, 4], 1, DASI, [512]] # 11-P4/16    #  40 40

  - [ 11, 1, Conv, [ 256, 3, 2 ] ] # 12-P5/32 # 256 20 20
  - [ [ -1, 10], 1, Concat, [ 1 ] ]  #13     #  20 20
  - [ -1, 3, C2f, [ 512 ] ]  # 14-P5/32  #512 20 20

  - [11, 1, nn.Upsample, [None, 2, "nearest"]] #15   #512 80 80
  - [[-1, 4], 1, Concat, [1]] # cat backbone #16  #80 80
  - [-1, 3, C2f, [512]] # 17                     #512 80 80

    #512 20 20    40 40    256 80 80
  - [[14, 11, 17], 1, DASI, [512]] # 18-P4/16  #  40 40

  - [18, 1, SCDown, [256, 3, 2]] # 19-P5/32 #  256 20 20
  - [[12, 19, 14], 1, Concat, [1]]     #20   20 20
  - [-1, 3, C2fCIB, [1024, True, True]] # 21 (P5/32-large)

  - [18, 1, nn.Upsample, [None, 2, 'nearest']] # 22-P3/8  #  80 80
  - [[15, 22, 17], 1, Concat, [1]]      #23         # 80 80
  - [-1, 3, C2fCIB, [1024, True, True]] # 24 (P5/32-large)

  - [[24, 18, 21], 1, v10Detect, [nc]] # Detect(P3, P4, P5)